Design of a microfluidics device to rapidly genetically analyze bronchial epithelial
cells for EGFR mutation
Emily Boggs
INTRODUCTION
Lung cancer is a growing epidemic in America. In 2012 alone, an estimate 226,000 people were
diagnosed with lung or bronchial cancer, while another 160,000 died [1]. In addition, lung cancer is the
second most commonly diagnosed cancer for men and women, just behind prostate and breast cancers,
respectively. More people, of both sexes, died from lung cancer in 2012 than any other type of cancer.
In fact, lung cancer killed more people than prostate, breast, pancreas, and colon cancers put together
(Figure 1).
Figure 1: Lung cancer deaths compared to death from other types of cancer [2].
Currently, the only definitive detection methods for lung cancer are chest x-rays, computed
tomography (CT), and surgical biopsy of possibly cancerous tissue. However, lung cancer is first
suspected based on the symptoms of chest pain, blood in expectorate, shortness of breath, and weight
loss [3]. Currently, the American Society of Clinical Oncologists suggests yearly screenings of low-dose
computed tomography for at risk individuals. The Society does not recommend, however, the screening
of individuals who may develop lung cancer from working with industrial carcinogens (like asbestos) or
individuals who have had significant exposure to secondhand smoke over the course of their lifetime. In
2012, these “never smokers” represented 10-20% of all lung cancer diagnoses [4].
No effective screening method for lung cancer exists apart from CT. CT can give detailed
anatomical information, but often lacks in the ability to differentiate between malignant and benign
cancerous lesions [5]. Chest x-rays are not sensitive enough to pick up on the earlier stages of lung
cancer [5]. In addition, both these methods are impractical for use as annual screening methods over a
number of years due to the high dosage of radiation that can be incurred on the patient. As a result, lung
cancer is not often definitively diagnosed until histology can be performed on tissue from a surgical
biopsy. However, by the time a surgical biopsy is performed, lung cancer is already suspected and may
be at an advanced stage. This has led researchers to look for alternative ways to diagnose lung cancer
using molecular biology and non-invasive sample collection methods.
Epidermal growth factor recpetors
Epidermal growth factor receptor (EGFR) is a protein found on cell membranes. Upon interaction
with growth factor ligands, the receptor activates several cellular signaling cascades, including ones for
cell division [6]. It should be no surprise then that mutations affecting the regulation of EGFRs have been
found to be associated with cancer [7]. Mutations often occur on exons 18-21 of the EGFR gene; this
string of code is associated with the part of an EGFR that increases kinase activity. Commons mutations
are a deletion in exon 19 and a point mutation at exon 21. These mutations are also extremely common
in “never smokers” who still develop lung cancer [8].
EGFRs are present in many cells, including bronchial epithelial cells where they have been
implicated in damage repair and mucus secretion [9]. Genetic analysis of these cells taken from the
airway can provide information regarding the presence of EGFR mutations.
Collection and enrichment of sputum
A sputum sample for genetic analysis can be obtained non-invasively [10] but requires
enrichment before genetic examination [11, 12]. After enrichment, each sample contains enough material
to undergo gene chip analysis as demonstrated by Jiang et. al [5]. Since the device presented here will
require approximately the same amount of genetic material as needed for a gene chip, a post-enrichment
sputum sample can be used for analysis.
Detection of EGFR for lung cancer diagnosis
Extensive research has already been completed on using genetic material to determine the
presence of mutated EGFRs [13]. In addition, The most commonly used method is direct sequencing, in
which DNA containing the mutated exon 19 are amplified and detected using gel electrophoresis.
However, these methods usually require large amounts of DNA in the form of samples from invasive
procedures, like tissue from surgical biopsy or pleural effusion from thoracentesis. Some procedures, like
direct sequencing, require that at least 20% of the cells examined must have the EGFR mutation for the
test to register as positive. Since the range of actual tumor cells containing the mutation in a histological
sample can vary from 5 to 100% [13], tests like direct sequencing can miss important diagnostic
information.
Another method for the detection of mutation-bearing genetic sequences is denaturing high
performance liquid chromatography (dHPLC). Sueoka et al. [14] determined that using dHPLC for
analysis was faster (five hours for dHPLC versus eighteen hours for direct sequencing) and cheaper
(direct sequencing was eight times more expensive). In addition, dHPLC provided more accurate results
than direct sequencing. Currently, HPLC (though not dHPLC) has shown promise as an analysis method
that can be miniaturized onto a microfluidics device [15]. Other screening methods for the EGFR mutation
include massively parallel sequencing analysis (or second generation sequencing) [16] and high
resolution melting [17]. However, all of these analysis methods require amplification of the genetic
material, a time-consuming process, and the latter two methods have yet to be studied for miniaturization
onto microfluidics devices. In addition, the need for an amplification step also requires that the sample
already contains high amounts of mutation-bearing cells; these samples will have to come from invasive
biopsy procedures like those mentioned above. Another, though much less commonly used, method for
EGFR detection is the use of immunochemistry to bind directly to the mutated receptors themselves [18].
However, specific recombinant antibodies are needed to prevent non-specific binding with wild-type EGF
receptors and other proteins [19].
The goals of the device presented here are to create a fast, reliable, inexpensive, and noninvasive method to screen at-risk individuals with lung cancer. Since the device is miniaturized onto a
microfluidics device, analysis time, reagent cost, and required sample volume will be much less compared
to current sequencing technologies. Since the device will require only a non-invasive sputum sample, it
can be used feasibly on at-risk populations instead of those already presenting symptoms of lung cancer.
In addition, the device will be able to detect EGFR mutations commonly present in non-smoking
individuals.
Benefits of using a microfluidics device
Microfluidic devices, and specifically lab-on-a-chip devices like the one proposed here, incur
several advantages over standard macroscale benchtop procedures simply due to their size [20]. The
most often cited benefit is that smaller sample sizes are needed. Microfluidic analysis is usually
inexpensive because the amount of possibly expensive reagents used is drastically reduced. In the realm
of clinical diagnostics, this translates to a reduction in sample size. This allows for less invasive
procedures, like sputum procurement, to be used diagnostically instead of surgical biopsy.
Because volumes are so small, diffusion is often the major drive for mixing samples. This
decreases the amount of time required for reactions or other processes to take place. With smaller
volumes and shorter data collection times, assays are more sensitive and results are more accurate.
METHODS
The methods used by the device for detection of a mutated mRNA sequence for EGFR are
presented below. The main focus is on detection of the mutated mRNA; mRNA is used instead of the
original mutated DNA as it demonstrates active translation of the affected receptor by the cell. After the
sample of bronchial endothelial cells in PBS is introduced into the device, the cells undergo
electrochemical lysis. The mRNA contained in the cells is then purified using a silica bead microcolumn.
The target mutated mRNA in the elution are then hybridized with complementary ssDNA (single-stranded
DNA), creating dsDNA (double-stranded DNA). The presence of either unhybridized ssDNA or hybridized
dsDNA changes the color of a colloidal gold nanoparticle solution, based on electrostatic interactions. A
blue color, which can be determined visibly, represents the presence of the mutation while a pink color is
represents a negative result.
Electrochemical Cell Lysis
Di Carlo et al. [21] have designed a novel reagent-free cell lysis microfluidics device. The device
uses tiny platinum electrodes to generate –OH and H+ ions in solution. The introduction of a high amount
of hydroxide ions will destroy cell membranes (Figure 2), allowing for the release of genetic material. A
filter (Figure 3) separates the cathode, which produces hydroxide ions, and the anode, which produces
hydrogen ions. The purpose of the filter is to keep the introduced cells in close proximity to the cathode
until lysis occurs; as a result, the filter size must be less than 10 microns, as this is the approximate
diameter of bronchial epithelial cells [22]. The hydroxide-containing lysate will be free to move pass the
filter and combine with the hydrogen ions being produced at the anode, creating a neutral solution for use
in the next step. This process has been shown to work with red blood cells, HeLa, and Chinese Hamster
Ovary cells, so testing will be required to prove effectiveness at lysing human epithelial bronchial cells. It
can be hypothesized that since the HeLa cell line comes from human cervical epithelial cells, the process
that works for HeLa may also work for epithelial bronchial cells.
Figure 2: The reaction responsible for lysing cell
membranes at the cathode [21].
Figure 3: A schematic and micrograph of the filter that
separates the anode from the cathode. Only lysed
material can pass through [21].
mRNA microcolumn Purification
Poeckh et. al [23] present a miniaturized silica bead microfluidics device for purifying DNA and
RNA from cell lysate. Their approach to the design mimics macroscale purification methods. The cell
lysate is pushed through a bed of silica beads in a solution with a high ionic strength, like sodium chloride
or guanidinium thiocyanate (GuSCN). The solution causes DNA and RNA to bind to the surface of the
silica beads. A subsequent wash through of a neutral solution removes all other non-bonded debris. The
DNA or RNA is then released from the silica beads with an elution buffer and pushed out the column.
Macroscale purification makes use of centrifugation to pull the lysate, wash, and elution solutions through
the column; however, this is not feasible for a microfluidics device. Instead, Poeckh et. al found that
using a pressure gradient to drive the fluid flow in the device was sufficient.
Poeckh et. al also found the elution times for RNA and DNA at a flow rate of 100 µL/hour with 6M
GuSCN at a pH of 8; most of the RNA elutes with the first 2 µL of solution, while DNA elutes between 3 to
6 µL. The solution used for the elution step was molecular biology grade water.
For the microfluidics device presented here, the mRNA purification step represents the most
complicated part of the setup, design-wise. Separate inlets would need to be addressed for the
introduction of GuSCN, wash buffer, and MBG water, as well as separate outlets for the removal of lysate
and elution. However, the volume of the reagents required are very small and could conceivably come
pre-packaged in the device.
Hybridization of mRNA to complementary ssDNA
The next step is to promote the hybridization of the purified mRNA to complementary ssDNA
bearing the mutation. The sDNA will have to be synthesized via RT-PCR from mRNA known to carry the
mutation [24]; this step will be carried out beforehand and separately from the rest of the procedures
described here, and will not involve the device. In the device, the purified mRNA and ssDNA will be
combined in a new chamber containing 10 mM PBS and 0.3 NaCl. The hybridization reaction will take
place over the course of five minutes at room temperature [25].
Li et. al [25] have also found a way to prevent the detection of mismatched hybrids (or ds’DNA).
Dissociation of the hybrid occurs after several minutes, by dissociation of ds’DNA occurs more quickly
than dsDNA. Waiting an extra two minutes after the hybridization reaction allows for dissociation of the
ds’DNA while the dsDNA remains bound.
If the mutated mRNA was present in the purified sample, dsDNA will be created during the
hybridization reaction. If no mutated mRNA was present, the other genetic material present in the elution
will fail to bind to the complementary ssDNA. As a result, a sample lacking the mutated mRNA will
produce only ssDNA after the hybridization step.
Addition of colloidal gold solution
The transduction method of the device relies solely on whether ssDNA or the hybridized dsDNA
is present in the sample. After hybridization and dissociation of any ds’DNA, the colloidal gold
nanoparticle solution is added to the chamber. As Li et. al [25] have discovered, ssDNA and dsDNA have
different electrostatic interactions with gold nanoparticles. The ssDNA can present a side of unbound
nucleotides; this causes the ssDNA molecules to stick electrostatically to the nanoparticles. In contrast,
the dsDNA is encased in a negatively charged phosphate-sugar backbone. As a result, few dsDNA
molecules will attach electrostatically to the nanoparticles.
The addition of a salt solution to a gold nanoparticle colloid causes the particles to aggregate; as
the color of gold colloidal solutions is dependent on particle size, this aggregation changes the color of
the solution. However, ssDNA electrostatically attracted to the gold nanoparticles act to negate the
attractive forces presented by the salt solution. As a result, a colloidal gold solution with ssDNA will not
change color upon the addition of salt solution. However, no such electrostatic barriers exist for the
particles if only dsDNA is present in the sample, and the solution will change color upon the addition of
salt solution (Figure 4). This change in color which is visible in 5 µL down to 100 fmol of DNA present in
the sample can be used to detect the presence of the EGFR mutation without the use of optical
instrumentation.
Figure 4: Color change associated with electrostatic interactions of gold nanoparticles with DNA [25].
TESTING
The steps taken by the device to diagnose the presence of the EGFR mutation are inherently
complex. Thus, testing of the device will begin with an analysis of each step individually.
Electrochemical cell lysis
The question of whether the electrochemical method of lysing cells is effective can be confirmed
via microscopic visualization. Di Carlo et. al [21] stained their HeLa cells with calcein to confirm cell lysis
in real time using a fluorescent microscope.
The next question is whether the lysed cells present in the sample can contribute a sufficient
quantity of genetic material to the analyte. The amount of EGFR mutated mRNA can be quantified by
running qPCR on the lysate. The colloidal gold solution requires at least 100 fmol of genetic material to
be present to allow for visualization of the color change. The quantity given by qPCR can also be used to
determine the amount of complementary ssDNA that needs to be added in the hybridization step.
mRNA microcolumn purification
The purification step needs to be tested to ensure that only and all genetic material from the
lysate passes through the silica bead column. Lysate known to contain the mutated mRNA will be
passed through and analysis of the elution will be performed using qPCR. This will also quantify the
amount of mutated mRNA and ensure that the concentration has not decreased below 100 fmol before
hybridization.
Hybridization of mRNA to complementary ssDNA
Northern blotting can be used to verify and optimize the hybridization the mRNA and DNA under
microfluidic conditions.
The main goal in testing the hybridization step is to optimize the amount of ssDNA that needs to be
added. Microarrays of oligonucleotides complementary to the ssDNA can be used. By using a
microarray with a number of oligonucleotides comparable to the concentration of mutated mRNA present
after elution, different concentrations of fluorescently-labelled ssDNA can be introduced to the
microarrays. The concentrations of the ssDNA added will be increased until there is no longer a
measurable increase in fluorescence. The highest concentration up to this point will be used for the
microdevice. Goff et. al [26] used a similar method of optimization for detection of miRNA.
Addition of colloidal gold solution
The color change associated with the gold colloid will have to be verified first with no genetic
material present. Next, different concentrations of the colloid will be added to the expected sample
concentrations of ssDNA and dsDNA, as determined above. The concentration that provides the most
definite color change will be used.
Once testing of all the separate components is completed, the sample will be run from start to
finish with all the concentrations and volumes expected to be used in the microfluidics device, and any
overall design flaws or changes can be made to the experimental protocols. Next, based on the
experimental protocols, the actual design of the device can begin. Different designs will be fabricated and
tested with sputum samples known to contain the EGFR receptor mutation, and samples that do not.
When a design has been chosen and optimized, clinical testing on lung cancer patients can
begin. The device will first be used alongside more conventional methods of detecting lung cancer.
Finally, a long-term study will be performed in which individuals testing positive on the device for the
EGFR mutation will be monitored for the possible development of lung cancer.
REFERENCES
[1] American Cancer Society. Cancer Facts & Figures 2012. Atlanta: American Cancer Society; 2012.
[2] American Lung Association. (2012). Lung cancer fact sheet. Retrieved from http://www.lung.org/lungdisease/lung-cancer/resources/facts-figures/lung-cancer-fact-sheet.html
[3] LungCancer.org. (2012). Screening for lung cancer. Retrieved from
http://www.lungcancer.org/find_information/publications/163-lung_cancer_101/274-screening
[4] McCarthy, W. J., Meza, R., Jeon, J., & Moolgavkar, S. H. (2012). Chapter 6: Lung cancer in never
smokers. Risk Analysis, 32(s1), S69-S84. doi:10.1111/j.1539-6924.2012.01768.x
[5] Jiang, F., Todd, N. W., Qiu, Q., Liu, Z., Katz, R. L., & Stass, S. A. (2009). Combined genetic analysis
of sputum and computed tomography for noninvasive diagnosis of non-small-cell lung cancer. Lung
Cancer, 66(1), 58-63. doi: http://dx.doi.org.ezproxy.ltu.edu:8080/10.1016/j.lungcan.2009.01.004
[6] Oda, K., Matsuoka, Y., Funahashi, A., & Kitano, H. (2005). A comprehensive pathway map of
epidermal growth factor receptor signaling. Molecular Systems Biology, 1, doi:
10.1038/msb4100014. Online.
[7] Lovly, C., Horn, L., & Pao, W. (2012, September 7). Egfr mutations in non-small cell lung cancer
(nsclc). Retrieved from http://www.mycancergenome.org/content/disease/lung-cancer/egfr
[8] Ren, J. H., He, W. S., Yan, G. L., Jin, M., Yang, K. Y., & Wu, G. (2012). Egfr mutations in non-smallcell lung cancer among smokers and non-smokers: a meta-analysis. Environmental and Molecular
Mutagenesis, 53(1), 78-82. doi: 10.1002/em.20680
[9] Burgel, P., & Nadel, J. A. (2004). Roles of epidermal growth factor receptor activation in epithelial cell
repair and mucin production in airway epithelium. Thorax, 59(11), 992-996. doi:
10.1136/thx.2003.018879
[10] Qiu, Q. et. al (2008). Magnetic enrichment of bronchial epithelial cells from sputum for lung cancer
diagnosis. Cancer Cytopathology, 114(4), 275-283. doi: 10.1002/cncr.23596
[11] Li, R. et. al (2007). Genetic deletions in sputum as diagnostic markers for early detection of stage i
non-small cell lung cancer. Clinical Cancer Research, 13, 482-487. doi: 10.1158/1078-0432.CCR06-1593
[12] Jiang, F., Todd, N. W., Li, R., Zhang, H., Fang, H. B., & Stass, S. A. (2010). A panel of sputum-based
genomic marker for early detection of lung cancer. Cancer Prevention Research, 3, 1571-1578.
[13] Ellison, G., Zhu, G., Moulis, A., Dearden, S., Speake, G., & McCormack, R. (2013). Egfr mutation
testing in lung cancer: a review of available methods and their use for analysis of tumor tissue and
cytology samples. Journal of Clinical Pathology, 66(2), 79-89. doi: 10.1136/jclinpath-2012-201194
[14] Sueoka, N. et. al (2007). Mutation profile of egfr gene detected by denaturing high-performance liquid
chromatography in japanese lung cancer patients. Journal of Cancer Research and Clinical
Oncology, 133(2), 93-102.
[15] Kim, J. Y., Cho, S. W., Kang, D. K., Edel, J. B., Chang, S. I., deMello, A. J., & O'Hare, D. (2012). Labchip hplc with integrated droplet-based microfluidics for separation and high frequency
compartmentalisation. Chemical Communications, 48(73), 9144-9146. doi: 10.1039/C2CC33774F
[16] Querings, S. et. al (2011). Benchmarking of mutation diagnostics in clinical lung cancer specimens.
PLOS ONE, 6(5), doi: 10.1371/journal.pone.0019601
[17] Borras, E. et. al (2011). Clinical pharmacogenomic testing of KRAS, BRAF, and EGFR mutations by
high resolution melting analysis and ultra-deep pyrosequencing. BMC Cancer, 11(406), doi:
10.1186/1471-2407-11-406
[18] Kawahara, A. et. al (2011). Identification of non-small-cell lung cancer with activating EGFR
mutations in malignant effusion and cerebrospinal fluid: rapid and sensitive detection of exon 19
deletion E746-A750 and exon 21 l858R mutation by immunocytochemistry. Lung Cancer, 74(1), 3540. doi: 10.1016/j.lungcan.2011.02.002
[19] Gupta, P., H., S. Y., Holgado-Madruga, M., Mitra, S. S., Li, G., Nitta, R. T., & Wong, A. J. (2010).
Development of an GFRvIII specific recombinant antibody. BMC Biotechnology, 10(72), doi:
10.1186/1472-6750-10-72
[20] van Dam, R. M. (2006) Solvent-resistant elastomeric microfluidic devices and applications.
Dissertation (Ph.D.), California Institute of Technology. http://resolver.caltech.edu/CaltechETD:etd12052005-234258
[21] Di Carlo, D., Ionescu-Zanetti, C., Zhang, Y., Hung, P., & Lee, L. P. (2004). On-chip cell lysis by local
hydroxide generation. Lab on a Chip, 5, 171-178. doi: 10.1039/b413139h
[22] Devalia, J. L., Sapsford, R. J., Wells, C. W., Richman, P., & Davies, R. J. (1990). Culture and
comparison of human bronchial and nasal epithelial cells in vitro. Respiratory Medicine, 84(4), 303312. [23] Poeckh, T., Lopez, S., Fuller, A. O., Solomon, M. J., & Larson, R. G. (2008). Adsorption
and elution characteristics of nucleic acids on silica surfaces and their use in designing a
miniaturized purification unit. Analytical Biochemistry, 373(2), 253-262. doi:
10.1016/j.ab.2007.10.026
[24] Life Technologies. (n.d.). Strategies for detecting mRNA Northern blotting, nuclease protection
assays, in situ hybridization, and RT-PCR. Retrieved from
http://www.invitrogen.com/site/us/en/home/References/Ambion-Tech-Support/northernanalysis/tech-notes/strategies-for-detecting-mrna.html
[25] Li, H., & Rothberh, L. (2004). Colorimetric detection of DNA sequences based on electrostatic
interactions with unmodified gold nanoparticles. Proceedings of the National Academy of Sciences
of the United States of America, 101(39), 14036-14039. doi: 10.1073/pnas.0406115101
[26] Goff, L. A., Yang, M., Bowers, J., Getts, R. C., Padgett, R. W., & Hart, R. P. (2005). Rational probe
optimization and enhanced detection strategy for microRNAs using microarrays. RNA Biology, 2(3),
93-100.